Newsgroups: comp.text.sgml Date: 01 May 1994 07:01:21 UT From: James Tauber \ Organization: The University of Western Australia Message-ID: <2pvk41$35c@styx.uwa.edu.au> Subject: Classical Greek Data Content Notation? I recently designed a classical Greek font that I'd like to make ISO compliant with regard to coding. I am aware of the Greek data content notation (where passages of Greek are tagged \) so far as it is a transliteration system for the actual Greek letters themselves but I'm interested if there is a standard for the diacritics, breathing marks, etc., when they appear in long passages of classical Greek. If there is an ISO standard for the Romanization of Classical Greek, would someone be so kind as to let me know how breathing, etc., is done? Thanks -- James K. Tauber, Undergraduate Student Centre for Linguistics, UWA, Australia E-mail: jtauber@tartarus.uwa.edu.au WWW: ftp://tartarus.uwa.edu.au/pub/jtauber/main.html Newsgroups: comp.text.sgml Date: 01 May 1994 15:13:50 UT From: Erik Naggum \ Organization: Naggum Software; +47 2295 0313 Message-ID: <19940501.2011@naggum.no> References: <24700@dog.cme.nist.gov> \ Subject: Re: sgml entities [Christoph Altenhofen] | (Hope you have misspelled the entity-call only in your posting. The | semicolon after \&Company.Name is missing, because it isn't followed by | a RE, but a comma.) note that the semi-colon is _not_ required when the character following the entity reference could not be part of the reference. the actual text of the standard reads [353:3]: The /refc/ or /RE/ can be omitted only if the reference is not followed by a character that could occur in the reference, or by a character that could be interpreted as the omitted _reference end_. \ -- Erik Naggum \ \ | memento, terrigena. ISO 8879 SGML, ISO 10744 HyTime, ISO 10646 UCS | memento, vita brevis. Newsgroups: comp.text.sgml Date: 01 May 1994 15:25:10 UT From: Erik Naggum \ Organization: Naggum Software; +47 2295 0313 Message-ID: <19940501.2012@naggum.no> References: <24700@dog.cme.nist.gov> <1994Apr29.220917.1558@xgml.com> Subject: Re: sgml entities [Eric R. Skinner] | If I may comment on the overall approach you are taking, rather than | the specifics of the SGML interpretation, I'd like to suggest that | using entities for boilerplate text is not the only technique | available. | | We'd argue that having to change your entity declarations each time | your boilerplate value changes is a bit of a hack, depending on the | scale of your system. it depends on how easily you can redefine entities. if you have to edit the document to do it, I would agree. this is your situation if you buy any of the major SGML parsers. if instead you use a parser that supports IMPLICIT LINK, you can override the entity in the link type declaration: \ \ ]> \ ]> \ This company, \&company, cannot be held responsible for anything. \ \ this could look like a prime candidate for SIMPLE LINK, but the complex wording in 12.1.4.3 appears to prohibit entity declarations in simple link type declarations subsets. this must be one of those minor mistakes that review processes are for. SIMPLE LINK is indeed simple, especially for the implementor, whereas IMPLICIT is somewhat harder. few have implemented either, so I suggest we allow SIMPLE LINK to behave as if it were better defined: \ ]> this would allow the same simple link to be used with a larger set of documents that all refer to the company by entity reference. this may be useful if you work for a company that changes its name frequently, such as many publishers. a further "complication" to some would be that LINKTYPEs need to occur between the document type declaration and document instance. this is a result of weak entity managers that are unable to extract substrings of files, but can also be rectified by smart parsers who, upon finding the end of the prolog, continue to read a set of LINKTYPE declarations from another file, conceptually part of the same entity. \ in any case, what we have here is an elegant mechanism to redefine entities according to processing needs, which also makes it possible to list all the relevant processing dependencies in the source document, to be selected at run-time by invoking the linktype. the use of a processing instruction to signal an error makes the default value hard to ignore. \ -- Erik Naggum \ \ | memento, terrigena. ISO 8879 SGML, ISO 10744 HyTime, ISO 10646 UCS | memento, vita brevis. Newsgroups: comp.text.sgml,comp.text Date: 01 May 1994 20:09:44 UT From: Gary Merrill \ Message-ID: <1994May1.200944.20383@ponds.uucp> References: <19940427.082224.679@almaden.ibm.com> Subject: Re: Single Text, Many Uses [Michael Priestley] | Actually, I think the chunks would end up being at the paragraph level | or even smaller; potentially even at the word level, as Gary Merrill | suggests, which could prove processor-intensive for anything but very | short documents. Given this, I think SGML has very little to add to | the process. It's still useful in other ways, of course, but most of | the SGML tags would be specific to a single medium, and so wouldn't | actually be "single-sourced" at all. For example: : | In these examples, the SGML tags are medium-specific, because the | organization of information is medium-specific. If any DBMS system can | format this on the fly, and display any particular medium or "version" | and allow me to change it (eg I change Chunk A in medium 1), and | automatically bring up Medium 2 at the place(s) where Chunk A is used | there, so I can assess the effect of my change in its various contexts, | I'd find it darned useful. If it operates on OS/2, I'll buy it now! | | But I don't think this is a solution, in itself, that owes much to | SGML. Information is being single-sourced, but its structure isn't; | and as I understand it, SGML tags the structure of information. As has | been previously suggested, perhaps this thread doesn't belong on | comp.text.sgml anymore, and it should settle in comp.text. : | I hope I've clarified the idea. I think this is what Gary Merrill and | others have been discussing (I could be wrong), and their discussions | might make more sense to you now (if what I've said made any sense). I think that this is a good presentation of the sort of potential complexity and the possible pitfalls with which I was concerned. And again, I think it demonstrates the general and fundamental problems that are faced in this area -- ones that SGML will not solve simply by being SGML, and ones which likewise any *other* representation must also grapple with. Gary Merrill -- "No reading, no inquiry has yet been able to remove my difficulty, or give me satisfaction in a matter of such importance. Can I do better than propose the difficulty to the public, even though, perhaps, I have small hopes of obtaining a solution? We shall, at least, by this means, be sensible of our ignorance, if we do not augment our knowledge." (David Hume) Newsgroups: comp.text.sgml Date: 02 May 1994 08:48:24 UT From: Yue-ming Lan \ Organization: ITRC Message-ID: <2q2eoo$7tb@news.csie.nctu.edu.tw> Subject: SGML->LaTex->EPS Dear SGMLers, I was told to find a solution to convert SGML documents to EPS, I would like to ask dear netters, how is to best way to achieved this. My idea is converting SGML to LaTex and then to EPS. I was told there is an utility does this, where to find it? Can it process double byte character? (I mean Chinese or Japanese)? Is there a Japanese version of SGML2Latex converter? Many thanks in advance. Yue-ming Lan csymlan@saturn.yzit.edu.tw Newsgroups: comp.databases,comp.text.sgml,comp.infosystems,comp.databases.theory,comp.databases.object Date: 02 May 1994 12:53:47 UT From: Martin Josko \ Organization: Leibniz-Rechenzentrum, Muenchen (Germany) Message-ID: <1994May2.125347.14369@news.lrz-muenchen.de> Subject: SGML->DSSSL(or others)->SQL(or others)? I'm interested at anything concerning the conversion from SGML-Documents into database languages (e.g., described in SQL, OODB-DDLs?) with the help of any standards (e.g., DSSSL) or any software (e.g., SINDA) or any other theoretical approaches. Does somebody have some hints or comments to this theme? -- Martin Josko Technical University Munich TU-Muenchen Department of Mathematics DVS-Weihenstephan Statistics and Data Processing D-85350 Freising Freising Germany email: martin@pollux.edv.agrar.tu-muenchen.de phone: +49-(0)8161-71-4506 fax: +49-(0)8161-71-4409 Newsgroups: comp.text.sgml Date: 02 May 1994 17:20:03 UT From: Mary Kroeger \ Message-ID: <9405021720.AA02508@pws09.aisg.com> Subject: NC SGML-User's Group March 5th 1994, a small group of young, talented friends gathered to discuss the possibilities of forming an SGML-User's Group. We yearned to discover more about SGML-supported software, how other corporations use SGML, and how we could make a difference in the future develoment of SGML Standards. This group of friends, now officially, the NC SGML-User's Group meets once a month to exchange ideas, information and insights into not only the SGML world, but also HyTime, DSSSL, and MathPack. The NC SGML-User's Group consists of the most skilled and knowledgeable Document Analysts, SGML editors, SGML proofers, and Project Supervisors employed by Gateway Conversion Technologies as well as Technical Writers from other RTP companies. On May 7th, the NC SGML-User's Group would like to open its doors to all individuals interested in attending an SGML information exchange. This meeting's agenda will include presentations by the HyTime and MathPack committees, formed by members of the SGML-User's Group. The meeting will take place 7:00 p.m. at Gateway Conversion Technologies, located in Morrisville, North Carolina. Gateway Conversion Technologies provides companies with the latest technology in manipulating and communicating information electronically. Supported electronic formats include raster images, ASCII, SGML, PostScript, and publishing formats such as Interleaf, FrameMaker, WordPerfect, and Microsoft Word. For further information please call Mary Kroeger or Suzy Wharam at +1 919 319 6500. Or send e-mail to \ or \. We look forward to meeting you! -- Mary Kroeger mkroeger@aisg.com Document Analyst Gateway Conversion Technologies Newsgroups: comp.text.sgml Date: 02 May 1994 18:56:50 UT From: Andrew Webber \ Organization: Games by Mail, Ottawa, Ontario Message-ID: <840241720D5.61R.123516000@gbm.ocunix.on.ca> Subject: Corel Ventura SGML? I haven't seen this discussed here yet; does anyone know what the "SGML support" in the upcoming Corel Ventura 5 will include? The software won't be shipping for another couple of months, but the upgrade order form arrived on Friday. One of the utilities listed is "TagWrite" which I assume is related to SGML. I wonder whether it will be an "SGML-to-Ventura" process similar to Miles 33's approach, a "Ventura-to-SGML" conversion, or both. Andrew Newsgroups: comp.databases,comp.text.sgml,comp.infosystems,comp.databases.theory,comp.databases.object Date: 03 May 1994 15:53:46 UT From: Pekka Kilpelainen \ Organization: University of Waterloo Message-ID: \ References: <1994May2.125347.14369@news.lrz-muenchen.de> Subject: Re: SGML->DSSSL(or others)->SQL(or others)? [Martin Josko] | I'm interested at anything concerning the conversion from | SGML-Documents into database languages (e.g., described in SQL, | OODB-DDLs?) with the help of any standards (e.g., DSSSL) or any | software (e.g., SINDA) or any other theoretical approaches. There is a paper entitled Christophides, Abiteboul, Cluet and Scholl: "From structured documents to novel query facilities" to be presented at the forthcoming PODS/SIGMOD'94. From the abstract: "This paper describes a mapping from SGML documents into an OODB and a formal extension of an OODB query language in order to deal with SGML documents storage and retrieval." The authors' e-mail address: {Vassilis.Christophides, Serge.Abiteboul, Sophie.Cluet, Michael.Scholl}@inria.fr. -- Pekka Kilpelainen Univ. of Waterloo, Dept. of Computer Science pkilpela@watsol.uwaterloo.ca (519) 885-1211 x6229 Newsgroups: comp.text.sgml Date: 03 May 1994 20:13:02 UT From: Michael Will \ Message-ID: \ References: <767060935.AA00364@bcnz.ac.nz> Subject: Re: Single Text, Many Uses (was Re: Real, true uses of SGML) [Errol Thompson] | I beleive that the traditional prose for printed material is not a good | approach to implementing online information. However there are | techniques for analysing information that lead to different styles of | presentation in both the printed and online media. One such method is | Information Mapping. : | What I am endeavouring to say is that the approach to analysing the | information becomes more critical when you shift from print media to | online but that a good analysis approach such as Information Mapping | may resolve those difficulties. Could you provide reference(s) to "Information Mapping" analysis technique? Thanks in advance. Newsgroups: comp.text.sgml Date: 04 May 1994 05:13:09 UT From: Sam Stevens \ Organization: Gateway Conversion Technologies Message-ID: <2q7at5$bp3@netprod1.aisg.com> Subject: FOSI FOSI? Wasn't he that guy on "Happy Days" with the leather jacket and the attitude? No? Well, you'd better start at the beginning. As a book designer, I have always provided my clients with a SPEC SHEET. Now they want a FOSI. So if you have _any_ info, your karma would benefit by your sending it (the information) to me. Thanks. -- =========Sam Stevens==sstevens@aisg.com================================================ I do not represent Gateway. I'm not sure I'm even allowed in that part of the building. ======================================================================================= Newsgroups: comp.text.sgml Date: 04 May 1994 09:14:00 UT From: Peter Bergstrom \ Message-ID: <2DC7EB42@noak.vxo.telub.se> Subject: Again: HyTime links in practice Hi, once again! Since my first posting on this subject, I've been re-writing much of the HyTime code for FMV Grund-DTD, due to Eliot Kimbers very useful reply. I have also gone through the HyTime standard in detail once more, and sorted a lot of things out. I'm sorry it took me so long to get back, but it took some time... I've also tried to follow the other threads that sprung up from my post, and it has been very interesting. Basically, I understand Carla's, Gary's and others standpoints, and I'm not fanatic about what we are trying to do (I hope). Eliot Kimber pointed out that sometimes compromises have to be done, and gave the IBMIDDoc as an example of a good compromise, and I believe we're trying to solve basically the same problem here. I also believe Wayne Wohler made a very good conclusion, and gave us all some things to think about. However, I'd like to go back to Eliots first reply and try to clarify a few things, and ask some new questions, still believing that I'm not the only one struggling with the HyTime standard and it's sometimes cryptic explanations. [Eliot Kimber] | In a HyTime application, elements that don't have any specific HyTime | function are normally declared as either form "HyBrid" or form | "sHyTime". HyBrid-form elements are simply able to contain other | HyTime forms, while sHyTime-form (suppress HyTime) elements suppress | any HyTime processing within them (note that the form sHyTime is a | renaming of the form nHyTime as defined in the Catalog of HyTime | Architectural Forms, available on the SGML Archiver server). Which form should I use, nHyTime or sHyTime? Is it an error to use one or the other? What is used by for example HyMinder? [Regarding elements "include-data", "include-nonSGML" and "reference":] | I may have missed something, but it's not clear how this element serves | to include things, unless it's to cause the inclusion of the element | that contains it (in which case, why not point to that element | directly). : | I would expect this element to have an ENTITY attribute that names the | data entity actually being included. : | I wonder if the descriptions of the anchors haven't been confused with | the links to those anchors, hence my confusion? I'd better clarify: Those three elements are used *only* when you want to refer to or include something. They are simply HyTime anchors for an independent link, and they are always the source anchor. These elements are located at the exact point of reference or inclusion, just as if a context link (clink) would have been used instead. We don't want the linking information in the modules, and therefore don't use clinks at all. You are right, they don't *do* anything! Their names are maybe misleading to you and others who understands the underlaying syntax, but we believe the names are useful to authors (but the SGML comments I made for these elements were incorrect, since they were directed to authors, not SGML/HyTime specialists). I think I'll explain this even more, since I don't believe it was very clear in my earlier posting. We envisage placing all links in one, and only one, module for the entire system. In all information modules, there will only be anchors, and they will know nothing about links at all. The "link module" will always act as the hub document, since it contains all links (and entity declarations for that matter). Even if I only want one single module I need to start from the link module in order to link in graphics, tool names, etc, which are all linked through HyTime independent links, found in the link module, i.e., the hub document. One could argue that we could use several link modules, maybe even one for each information module, but we see a very great benefit of using only one link module: The link module could, in an application such as the FMV storage system, be modeled with STEP/Express and converted into a data base, i.e., a good part of the data base structure would be given, and we would have a good start for a document management system. The HyTime syntax, and the link module, would be created from this data base at export. They are only necessary for interchange. And then, they describe all links between the modules and elements in a unambiguous way by the use of Hytime. Of course, an application could also be built around the link module, and use a HyTime engine to manage all links. Another benefit of using one single link module is that we don't have to duplicate Entity references and namelocs/nmlists pointing at the same module/ information. We understand that this concept might not be the best for the production of data, but we are designing a storage DTD, and there is nothing that stops producers from creating the information in accordance with a modified DTD that allows HyTime links inside information modules, if only they deliver in according to this DTD. [Regarding attribute "datatype" on the endterm element:] | Nice use of endterm. Note that the datatype attribute should be | unneeded because the object located will indicate its type as part of | the data entity declaration. By the same token, elements carry their | own type information, which the processing application doesn't need | until it actually gets the data (it doesn't need formatting information | before it has the stuff it's going to format). I didn't explain the use of "datatype" clear enough, I'm sorry. In the information modules, we don't have any elements named \, \, \ or alike, where the user have to enclose the name of the tool/sparepart/consumable with the appropriate element. Instead, we want all tool names, sparepart names and id's, etc., to be included into the module from a common storage, e.g., a "tool-module" or an LSAR data base. This is done by using the anchor element "include-data", which is only used for this purpose. There are, especially among consumables, frequent changes of the names and/or id's. Today, references to consumables are often made to the safety precautions document, rather than to the FMV id for it, just because this FMV id changes with every brand of kerosine, or whatever. What we want to achieve by this is that you link to a generic tool (or specific, depending on your needs at that specific point in your module), and whenever the id or name changes, the change is made in one place. Now, if one anchor point is \, what is the other? It could be a bibliographic reference, since the object is a real object that is not stored in the computer. This would be a logical solution, but not very efficient (or?). We envisage the other anchor as being one element in a "material-module" or "techdata-module". That element have a sub-element structure, which includes a lot of information regarding the tool, consumable, etc. My problem here is that I want to select only some of that sub-element structure, and I don't know what parts of it I'm interested in when writing the module. Let me explain: In the infomodule I have this: Clean the parts carefully, using \. In the "techdata-module" we have the following: \ \ \Kerosene \ \ \ \ \ \X12-Kerosene \ \F1234-567890 \ \KEM887-7654H-W \ \NOK56776-TY643 \ \ \ \ \ \ \ Now, in the printed document I don't want all different ID's to show up, just the name and the reference to the safety precautions document (must be retrieved from the safetyprec link, but that's another story). In a screen application, say, I want only the name as a "button". When I place the cursor above the button, I want a pop-up that shows the FMV-ID for all alternatives, and when I click the button I want to traverse to the safety precautions document. OK, so what about attribute "datatype", then? Basically it is up to the application to decide what information should be presented when an inclusion of data is made, as above, but we feel that we ought to give the author the possibility to *suggest* what information should be presented. If a second reference to kerosine is made three lines further down, it might not be necessary to present as much information as the first time, as an example. The attribute "datatype" is declared to give that possibility of suggesting how presentation of the data is going to be made. It's a CDATA attribute, but with (so far only implied) lextype=GI+. The GI's should also be one that exists inside the referenced element (e.g., "name" in the example above). (By the way, the name datatype is awful and doesn't say what it does. It will be changed.) I hope this long explanation made some sense... I'm wondering if this is the same thing as the HyTime attribute "reftype" attempts to do? I don't think I can simply change to reftype here, but maybe reftype will do the job better for me? | \ | \ \ In the first example the public id is declared as a DTD, in the second as a notation. Is it even possible to define a Notation as a DTD? Why not as an SGML document, subdocument, or text entity, then? What are the implications of the two different ways to declare the public entity? Any recommendations? My guess is that NOTATION must be used, at least when you don't know which DTD the module (document) uses. | Note also that issues of document fragment re-use can be solved by | using subdocuments as the re-usable components of larger documents (I | explain this use of subdocuments in my forthcoming book, and you can | also find an earlier post of mine on the subject in the SGML archive). I found an earlier posting of yours, a reply to Robin Cover entitled "Re: SUBDOC and (SGML) ESIS". What you say there is very interesting, and I hadn't thought of the possible name conflicts with just including elements into modules. But further questions arise: - Is it possible to link to an element inside an SGML entity as a SUBDOC (e.g. an element inside a module, not the entity as a whole)? We want to do this to create variants, for example. - How does HyTime unified name space help if the subdoc becomes part of a larger document? And how can a conref attribute have the same value whether the content is located in the same SGML entity (and name space) or if it is an external entity? In the first case is simply an IDREF, in the second it's an ENTITY. I can't even see that we can use the same attribute declaration. | Hope you find my comments useful. You've helped my point exposing a | hole I had in my book (no discussion of multiple locations). I certainly do! I was hoping to get a response like this, but I didn't dare believe in it. I'm happy that my posting helped at your side, too. That, I hadn't expected! Peter -- Peter Bergstr\öm internet: pebe@telub.se Telub Inforum AB phone: +46 - 470 - 425 97 351 80 V\ÄXJ\Ö fax: +46 - 470 - 397 96 SWEDEN Newsgroups: comp.text.sgml Date: 04 May 1994 14:06:00 UT From: Lori Snyder \ Message-ID: \ Subject: CONREF I have a question about the default value CONREF. I know that if the attribute is specified, the element associated with it becomes an empty element. If a value is not specified, then the element will contain whatever its content model is. What I haven't figured out is the minimiztion of the end tag for the element. If I do not have minimization on the end tag, the 'SGMLS' parser gives me a warning. I've been told that Author/Editor and XGML Validator do not issue a warning. I didn't see anything about the minimization for this particular case in the SGML books I have. Can anyone enlighten me on this issue? \ -- Lori Snyder lsnyder@cs.uah.edu Huntsville, AL Newsgroups: comp.text.sgml Date: 04 May 1994 14:18:49 UT From: Chet Ensign \ Message-ID: <199405041420.AA24030@Norway.EU.net> Subject: Next meeting of the SGML Forum of NY The SGML Forum of New York invites you to attend "Max-Factors for Successful SGML" Date: Monday, May 9, 1994, 5:30 p.m. Place: McGraw-Hill, Inc. 50th Floor Executive dining rooms (elevators are across from the entrance) 1221 Avenue of the Americas (lobby entrance off 49th Street) Agenda: 5:30 Registration and Opening Remarks 5:45 Keynote Speaker: Robert J. Glushko, Passage Systems, Inc. "Making a successful migration to SGML" 6:30 Refreshments 7:00 Vendor Demonstration: Robert J. Glushko, PassagePRO "A focus on tools alone is insufficient to ensure a successful project," states Bob Glushko; "What is needed for success with SGML is a framework in which technology and process fit together." At our May meeting, Bob will present the framework and the key factors that every successful SGML project must have. He will describe the principles he has adapted in the course of helping a number of companies make a successful migration to electronic document delivery and SGML. Bob says that he likes to make his talks "an antidote for the usual technology and syntax-intensive discussions." For our vendor demonstration, Bob will show us PassagePRO, an SGML-based document management and production system that supports the automated conversion of information into SGML and the subsequent assembly of SGML into hardcopy and online documents. Upcoming Programs: Our final meeting before the summer break will be held on June 14th in conjunction with the Graphic Communication Association's Online '94 conference. Evan Owens of the University of Chicago Press will tell us how they implemented SGML. David Silverman will talk about the services offered by Data Conversion Laboratory, a leading SGML conversion service bureau. And Bob Hecht and Sybil Parker of McGraw-Hill will describe the development of the Encyclopedia of Science and Technology, a multimedia encyclopedia produced using SGML. About the Forum The SGML Forum of New York is a nonprofit organization devoted to the exchange of ideas and information about SGML. Organized primarily as a user group, the Forum seeks to promote an understanding of the scope and benefits of the SGML standard and to further its practical application within a variety of industries, including publishing, financial services, insurance, pharmaceuticals, and telecommunications. Membership helps support the Forum's activities, but it is not required for attending general meetings. If you would like to receive word about future Forum meetings, send your name and address to: SGML Forum of New York, Inc. Bowling Green Station P.O. Box 803 New York, NY 10274-0803 We will be happy to put you on our mailing list. Best regards, /chet -- Chet Ensign +1 212 736 6250 X4349 Information Builders, Inc. internet: doccoe@ibivm.ibmmail.com 1250 Broadway ibmmail: USUBUVMV@IBMMAIL New York, NY 10001 compuserve: 73163,1414 Newsgroups: comp.text.sgml Date: 04 May 1994 17:29:35 UT From: Lloyd Rutledge \ Organization: UMass-Lowell Computer Science Message-ID: <2q8m1v$rdh@ulowell.uml.edu> References: <2DC7EB42@noak.vxo.telub.se> Keywords: HyTime, hyperlinking Subject: Re: Again: HyTime links in practice [Peter Bergstrom] | Which form should I use, nHyTime or sHyTime? Is it an error to use one or | the other? What is used by for example HyMinder? The architectural form "nHyTime" was specified in the HyTime Draft International Standard and was replaced by "sHyTime" in the final version of the standard. Thus, of course, you should use "sHyTime". | [ discusses "datatype" attribute which specifies GI of child of | identified element to present ] | I'm wondering if this is the same thing as the HyTime attribute | "reftype" attempts to do? I don't think I can simply change to reftype | here, but maybe reftype will do the job better for me? I don't think reftype is applicable here. The attribute reftype specifies what element types an UID reference is allowed to reference an instance of. It is a HyTime common attribute, and they typically serve as what could be loosely called "DTD extensions" -- that is, they specify what is allowed in a document in ways the go beyond the restrictions a DTD can place. Instead of stating what references are allowed, you want to specify as the final reference a particular child of the element directly referenced. I would suggest looking at the architectural forms in the location address module to do this trick. Perhaps you could use a treeloc or a pathloc to specify the child by the ordination of its appearence in the parent's content. Or even better, use a HyQ query to access the child of the referenced element that has that GI. -- Lloyd Rutledge Interactive Media Group Computer Science Department University of Massachusetts One University Avenue Lowell, MA 01854 USA voice: 508-934-3554 fax: 508-452-4298 email: lrutledg@cs.uml.edu Newsgroups: comp.text.sgml Date: 04 May 1994 17:36:54 UT From: David Megginson \ Organization: Department of English, University of Ottawa Message-ID: \ Keywords: entities pubid Summary: Are device-specific entities better than descriptive entities? Subject: Using Public Identifier Versions I have been doing some thinking about the version field in public identifiers, and I wonder whether it isn't best to avoid it altogether. For example, there are two models for handling the character entities when converting an SGML document to LaTeX: in the first, simply use the standard public identifiers like ISO 8879:1986//ENTITIES Added Latin 1//EN replacing the entities with SDATA sequences like "[aacute]", then in the postprocessor, convert those SDATA sequences to LaTeX sequences like "{\\'a}". The second approach is to make a LaTeX-specific version of the entities, declared with the public identifier ISO 8879:1986//ENTITIES Added Latin 1//EN//LaTeX where the entity file contains declarations like \ instead of \ (With SGMLS, I could set SGML_PATH to something including "/usr/lib/sgml/public/%O/%C/%V/%T", and keep the LaTeX-specific entities in "/usr/lib/sgml/public/ISO_8879:1986/entities/LaTeX/Added_Latin_1" -- I understand that the new version of SGMLS uses catalogues to simplify this somewhat.) I have read Goldfarb's brief comments on this point in the HANDBOOK, but he does not seem to express a qualitative judgement on the two approaches. The second seems good, since it allows the SGML parser itself to do most of the work, using the entity manager for its intended purpose. Personally, however, I think that the first is much better style -- granted that the postprocessor ends up doing a lot of work which the entity manager could have handled, but the document is no longer linked to a single conversion format. If I use the public identifier "ISO 8879:1986//ENTITIES Added Latin 1//EN//LaTeX" in my document, I am declaring that this document will be used for conversion to LaTeX and nothing else; on the other hand, if I leave off the "//LaTeX", I have not committed myself, and can use the document for anything from LaTeX to database or hypertext applications. I would be interested to hear other people's views on this subject. David -- David Megginson Department of English, University of Ottawa, dmeggins@aix1.uottawa.ca Ottawa, Ontario, CANADA K1N 6N5 dmeggins@acadvm1.uottawa.ca Phone: +1 613 564 6850 (Office) ak117@freenet.carleton.ca +1 613 564 9175 (FAX) Newsgroups: comp.text.sgml Date: 04 May 1994 23:09:14 UT From: Erik Naggum \ Organization: Naggum Software; +47 2295 0313 Message-ID: <19940504.2068@naggum.no> References: \ Subject: Re: CONREF [Lori Snyder] | What I haven't figured out is the minimiztion of the end tag for the | element. If I do not have minimization on the end tag, the 'SGMLS' | parser gives me a warning. I've been told that Author/Editor and XGML | Validator do not issue a warning. | I didn't see anything about the minimization for this particular case | in the SGML books I have. Can anyone enlighten me on this issue? Goldfarb, [408:20] reads 11.2.2 Omitted Tag Minimization "O" should be specified for end-tag minimization if the element has a content reference attribute or a declared value of "EMPTY". Note -- Specifying "O" serves as a reminder that empty elements do not have end-tags (although this has nothing to do with markup minimization). I don't necessarily agree with this suggestion. if the end-tag is required to be absent, it is not "omissible", and it is in fact confusing to have it specified as such. marking it as omissible also means you can't enforce the end-tag when this is something you might specifically want to do. -- Erik Naggum \ \ | memento, terrigena. ISO 8879 SGML, ISO 10744 HyTime, ISO 10646 UCS | memento, vita brevis. Newsgroups: comp.text.sgml Date: 04 May 1994 23:19:08 UT From: Erik Naggum \ Organization: Naggum Software; +47 2295 0313 Message-ID: <19940504.2069@naggum.no> References: \ Subject: Re: Using Public Identifier Versions [David Megginson] | I have been doing some thinking about the version field in public | identifiers, and I wonder whether it isn't best to avoid it altogether. I think it should be implemented the right way, instead. I'd be surprised if you _did_ like the results you have got, considering that you do it the wrong way. | Personally, however, I think that the first is much better style -- | granted that the postprocessor ends up doing a lot of work which the | entity manager could have handled, but the document is no longer linked | to a single conversion format. If I use the public identifier "ISO | 8879:1986//ENTITIES Added Latin 1//EN//LaTeX" in my document, I am | declaring that this document will be used for conversion to LaTeX and | nothing else; on the other hand, if I leave off the "//LaTeX", I have | not committed myself, and can use the document for anything from LaTeX | to database or hypertext applications. good argument, and in fact arguing strongly in favor of what the standard suggests (ISO 10.2.2.5; Goldfarb [390:9]): When a system accesses public text for which a public text for which a _public text display version_ could have been specified but was not, it must substitute the best available system-dependent version for the display device in use. If there is none, no substitution occurs. that is, leave it out of the document, and leave it to the SGML parser to inform the entity manager to look for a public text display version. that is, you get _both_ of the good things you describe, and _none_ of the bad. so, DO NOT USE public text display version in the document. but DO USE IT in the catalog, and obtain an entity manager that can deal with this very useful aspect of character entity sets in SGML. those that cannot are very clearly deficient. there is not even a choice in the matter. the wording above is "it must substitute the best available system-dependent version". only if the SGML system itself is lacking such device-dependent entity sets can it be excused of this requirement, but that is doing its users a major disservice and forcing applications to do what the parser and the entity manager _should_ be able to do, just as you point out. however, all is not lost if you have a deficient entity manager, but you _do_ have system-dependent versions of entity sets. you can do the path hack or the catalog hack and somehow specify which path or catalog to use at runtime, but how about using LINK (provided you don't have a deficient parser, as well :-)? Link processing declarations can override entity declarations in the document type to which it refers, and you can override entity declarations to use the public text display version, under specific user control for the individual, _named_ link processing declaration. the name will have to be specified at run-time, but that's no different from specifying the display device to use, or the path or catalog to use. since the display device is known to the application, and can more easily be communicated to the parser than a link type with an unknown name (to the parser, that is), using the public text display version is very clearly the best of the available options. by the way, I think using the public text display version for other things than just "display versions" would be a tremendously good idea. just call it "version", and it may apply to document type declaration subsets that are public text, and exist in various versions, but none without a version. think of a NOTATION declaration which can be either a specification of the notation, or a program or program fragment that validates or interprets the notation. (NOTATION is currently excluded from the kinds of public text identifiers that can have it, but this can be fixed.) a link processing declaration subset may itself be device-dependent, and likely is, so it, too, may be subject to this versioning scheme, not just character entity sets. the possibilities of an _implemented_ versioning scheme like this are numerous, and the problems solved by it are so many and so obvious that one is led to wonder why vendors have not implemented it. my guess it that if you have do "conversion" from [aacute] to whatever you really need, you don't get around using their proprietary scripting language. that's probably why they haven't implemented LINK, either. only when enough people know about these possibilities can a need for them be felt in the market, and vendors be suggested to implemented what they otherwise wouldn't want to. -- Erik Naggum \ \ | memento, terrigena. ISO 8879 SGML, ISO 10744 HyTime, ISO 10646 UCS | memento, vita brevis. Newsgroups: comp.text.sgml Date: 05 May 1994 03:13:19 UT From: Kevin C Cole \ Message-ID: <2q9o8f$qqs@news.delphi.com> Keywords: sgml DOS Subject: DOS-based SGML formatters? Does anyone know of DOS-based software that will format SGML documents? I haven't seen any since I wrote some doc for IBM about 10 years ago and used their PC-GML formatter w/Sherpa. I'd really appreciate any information. Thanks. --Kevin kccole@delphi.com Newsgroups: comp.text.sgml Date: 05 May 1994 05:53:44 UT From: Martin Hollender \ Organization: Free University of Berlin, Germany Message-ID: \ References: <767060935.AA00364@bcnz.ac.nz> \ Subject: Re: Single Text, Many Uses (was Re: Real, true uses of SGML) [Michael Will] | Could you provide reference(s) to "Information Mapping" analysis | technique? Thanks in advance. Horn, R.E. (1989) Mapping Hypertext. The Lexington Institute, Lexington, MA Contains a lot of further references to "Information Mapping" Ciao Martin -- Martin Hollender Man-Machine Systems Lab University of Kassel D-34109 Kassel Germany tel. ++49-561-804-2701 email: mh@imat.maschinenbau.uni-kassel.de fax. ++49-561-804-3542 Newsgroups: comp.text.sgml Date: 05 May 1994 06:12:00 UT From: Peter Bergstrom \ Message-ID: <2DC91527@noak.vxo.telub.se> Subject: HyTime prolog Hi! Here's a few questions concerning the prolog of a HyTime document. As far as I can see from the HyTime standard, the APPINFO parameter should be: APPINFO "HyTime" But if I choose to rename (in my DTD) either the HyTime attribute itself, or the HyName attribute, then the mapping of these should be included, say: APPINFO "HyTime HyTime=My-HyTime-Name,My-HyName-Name" Is this correct? Can somebody verify or correct what I stated above? Regarding the HyTime Support Declarations (clause 6.3.2) I have the following thoughts at this time: Let's say I have the following Support Decl's: \ \ \ \ According to the standard they shall appear "in the document prolog, prior to the base document type declaration". I can then see three possibilities: 1. They can be placed as the first thing in the Declaration Subset of the DOCTYPE statement (which I believe is actually part of the base document type declaration): \ \ \ \ \ ]> 2. They can be placed in the SGML Declarations File, after the SGML Declaration, of which APPINFO is the last parameter. This is however a "grey zone" and I have no idea if any parser/HyTime engine will ever see the declarations here. \ \ \ \ \ 3. They can be the very first thing of the DTD, but it seems to me that then the "base document type declaration" has already started (which might be the case even for alternative 1 above). (I don't think it's necessary to give an example of this.) I find all these unsatisfactory, and I see no example of this in the standard. Are there other alternatives? How do other people do? What do HyTime engines expect (exisiting and in development)? Thanks! Peter -- Peter Bergstrom internet: pebe@telub.se Telub Inforum AB phone: +46 470 425 97 351 80 VAXJO fax: +46 470 397 96 SWEDEN Newsgroups: comp.text.sgml Date: 06 May 1994 00:05:02 UT From: Eliot Kimber \ Message-ID: <2qc1je$g4o@search01.news.aol.com> References: <2DC91527@noak.vxo.telub.se> Subject: Re: HyTime prolog [Peter Bergstrom] | 1. They can be placed as the first thing in the Declaration Subset of the | DOCTYPE statement (which I believe is actually part of the base | document type declaration): | | \ | \ | \ | \ | \ | ]> Your answer 1 is the only right answer. The intent for putting the declarations in the SGML document proper, and not in the public part of the DTD or in the SGML declaration itself, is that each instance may have different requirements, and specifically, because HyTime validation is done on an instance basis. Of course, there's no reason these PIs couldn't be put into an external parameter entity and referred to from within the DTD subset at the proper point -- it would be trivial to add the declaration and reference to a document template or something. I don't know if any current tools enforce this requirement. -- \

W. Eliot Kimber (kimber@passage.com) Systems Analyst and HyTime Consultant Passage Systems, Inc., 9971 Quail Blvd., Suite 903, Austin TX 78758 (512)339-1400 465 Fairchild Dr., Suite 201, Mountain View, CA 94043, (415) 390-0911 \

Newsgroups: comp.text.sgml Date: 06 May 1994 08:28:07 UT From: N F Drakos \ Message-ID: <1994May6.082807.19224@leeds.ac.uk> References: \ Subject: Re: Single Text, Many Uses (was Re: Real, true uses of SGML) [Errol Thompson] | I beleive that the traditional prose for printed material is not a good | approach to implementing online information. However there are | techniques for analysing information that lead to different styles of | presentation in both the printed and online media. One such method is | Information Mapping. Some of the issues in "single text, many uses" are discussed in a paper to be presented at WWW'94 entitled "From Text to Hypertext: A Post-Hoc Rationalisation of LaTeX2HTML". In particular it proposes ways to address common objections to automatic conversion (in the context of HTML) such as: o Word-processing formats contain presentation information that cannot be reproduced in HTML o Conventional paper-based documents have a linear structure and are therefore unsuitable for conversion to the hypertext form o It is difficult to determine the level of granularity for each hypertext node o Some documents require custom navigation aids o The conversion process is inflexible and converted documents cannot take full advantage of hypermedia capabilities o There are subtle differences between electronic and paper-based versions and this will compromise quality This paper is available online as: http://cbl.leeds.ac.uk/~nikos/doc/www94/www94.html This node contains a link to a postscript version. Also, the node http://cbl.leeds.ac.uk/~nikos/doc/www94/section3_5.html#refgat contains an interactive control panel which can be used to try out the conversion mechanism on different small examples and see the results. (You will need a forms-capable browser to use this - eg a recent version of Mosaic for X). Nikos. -- Nikos Drakos Computer Based Learning Unit nikos@cbl.leeds.ac.uk University of Leeds http://cbl.leeds.ac.uk/nikos/personal.html Newsgroups: comp.text.sgml Date: 06 May 1994 12:42:23 UT From: Frank Teusink \ Organization: CWI, Amsterdam Message-ID: \ Subject: Q: Is SGML what I'm looking for? Hi, I am looking for a way to transform a relational database of `famous persons' (result of a hobby of me and a few others) to something that is more flexible. Right now, the database contains for each person a number of data-fields (i.e., name, data-of-birth, place-of-birt, occupation, etc), and a text-field, plus a number of relations on persons (i.e., parent, marriage, etc). It is quite hard to generate readable `articles' from this database; - one has to generate a readable paragraph of text from the data-fields, and - combine it with the text-field to a readable `article'. Via the HTML language, I arrived at SGML, which might be a solution to my problem. Ideally, I'd like to transform the database to something like \ \

\Henry IV\ (Bolingbroke), \king of England\\

Son of \...\ and \...\. Born on \...\ at ... blabla ... blabla ... \ Here, `h1' would be a mark-up tag to define the layout of the article (there could be others too), while the other tags used in the example would be `attribute-tags'. However, I would still need some database-functionallity. Most importantly, I need to be able to generate indexes on things like \ and \, and be able to answer querie to the database. I have some other things on my wish-list, but I won't bother you with them righ now ;-) Is this possible within the framework of SGML? Are there `SGML-databases' that can be used for something like this? Where are they? As an aside, I would be grateful for references to a good book on SGML. Regards, Frank Teusink frankt@cwi.nl Newsgroups: comp.text.sgml Date: 06 May 1994 13:29:01 UT From: Satoshi Tsuchiya \ Organization: Sect. 5, Planning Dep., Program Products Group, Fujitsu Ltd. Message-ID: \ References: <2q2eoo$7tb@news.csie.nctu.edu.tw> Subject: Re: SGML->LaTeX->EPS [Yue-ming Lan] | My idea is converting SGML to LaTex and then to EPS. I was told there | is an utility does this, where to find it? Can it process double byte | character? (I mean Chinese or Japanese)? Is there a Japanese version | of SGML2Latex converter? I have modified "sgmls" (free sgml parser written by James Clark) to handle Microsoft Kanji code (known as Shift-JIS code). Unfortunately it is not double byte oriented, but you can convert SGML document which is encoded by Shift-JIS code to other tagged document according to a conversion specification. Speaking roughly, "sgmls" treats a double byte encoded character as two single byte encoded character internally. Because applications don't have to know about internal process of a parser, there is no practical problem. If you would use Extended UNIX code (EUC) system, "sgmls" can treat multibyte encoded character. In case of Microsoft Kanji code system, it uses ASCII graphic character area as a second byte of double byte code. So single byte oriented parser misunderstands some of second byte of double byte code as a significant SGML character. For this case, I have modified "sgmls" source to convert Microsoft Kanji code system to EUC-base code system at I/O calling point. If you are interested in my modification, please mail to me. -- TSUCHIYA Satoshi | tsuchiya@sysrap.cs.fujitsu.co.jp | NIFTY-ID:GDC02435 $@EZ(J $@20(J $@E/(J ($@O5<@$N?M(J) | $@IY;N(J$@DL(J Middleware $@;v6HK\\It(J $@4k2hIt(J $@7W2hIt(J Newsgroups: comp.text.sgml Date: 06 May 1994 18:47:02 UT From: Jeremy LDI \ Organization: America Online, Inc. (1-800-827-6364) Message-ID: <2qe3b6$f7@search01.news.aol.com> Subject: SGML to LATEX I would be interested in any work done in converting mathematical text from SGML to LATEX. Possible collaboration on developing a filter? Newsgroups: comp.text.sgml Date: 06 May 1994 21:39:55 UT From: Lori Snyder \ Organization: Computer Science Dept., Univ. of Alabama-Huntsville Message-ID: <9405062139.AA01735@laser13.cs.uah.edu> Subject: SGML Instance tagged to 87269 spec Lately there has been a lot of talk about software that is MIL-D-87269 compliant. There are vendors claiming that their product is FULLY compliant. They will demo their product, but they never show what the SGML instance looks like. Why is that? Does anyone know where I could get a SGML instance tagged to the 87269 spec? The spec itself is very complex and very hard to understand without an example of how some of the tags are to be applied. Take for example the "alert" tag. There are several attributes, but nothing says what the attribute values are supposed to mean. Does the alert tag have to appear in every frame that it applies to? How else would you know when the alert no longer applies? \ -- Lori Snyder lsnyder@cs.auh.edu Newsgroups: comp.text.sgml Date: 07 May 1994 02:14:28 UT From: Kay Kit Chan \ Organization: Information Technology Institute, National Computer Board, S'pore Message-ID: \ Subject: Fax/Email address of Interleaf Hi, From a product brochure, I managed to obtain the address and telephone number of Interleaf as: Prospect Place 9 Hillside Ave Waltham, Massachusetts 02154 Tel: +1 617 290 0710 but can't seem to find their fax or email address. Can someone help? Please direct replies to the email address below. Thanks in advance! --- Kay-Kit CHAN | Internet: kaykit@iti.gov.sg Information Technology Institute | Tel: +1 65 772 0920 National Computer Board of Singapore | Fax: +1 65 777 3043 Newsgroups: comp.text.sgml Date: 07 May 1994 08:39:34 UT From: Hans Holger Rath \ Organization: Computer Graphics Center, Darmstadt, Germany Message-ID: \ Subject: [Q] DTD of the chemical industry? Hi everybody! A friend of mine asked me to post this article into comp.text.sgml because he has no news access. So, please send any answers to him via email. His address is: Markus Diermann \ ------------------------ Is there a group that has developed a DTD for the chemical industry? Any hints are useful! Thanks, --Markus ------------------------ Newsgroups: comp.text.sgml Date: 07 May 1994 15:22:58 UT From: Mark Leisher \ Organization: Computing Research Lab Message-ID: \ Subject: FOSI info This is probably a FAQ somewhere, but I couldn't find it: where can I get info on FOSI other than the files at ftp://wpaftb1.wpafb.af.mil/sgml/? Being a novice, these files are a little daunting. -- mleisher@crl.nmsu.edu Mark Leisher Computing Research Lab "Elegance is not optional!" New Mexico State University -- Attributed to Richard O'Keefe Box 3CRL, Las Cruces, NM 88003 Newsgroups: comp.text.sgml Date: 07 May 1994 17:56:32 UT From: Bill Krewin \ Organization: Regulation Scanning Message-ID: \ Keywords: sgml consultant Subject: consultant needed Our company publishes information electronically. We have approx 100,000 documents that are tagged in a proprietary format. Other companies have expressed strong interest in purchasing this data and accessing it via various search software. The search software these other companies use is typically a combination of full text and hypertext. We want to convert from this proprietary format to SGML. Our experience with SGML is that we can spell it. We have a technically competent staff of four programmers and a data operations department of twenty people. What we need is an experienced guide that will assist us in our journey. Please reply via email if you are interested or can assist us in locating such an individual. Newsgroups: comp.text.sgml Date: 08 May 1994 14:32:00 UT From: Peter Bergstrom \ Message-ID: <2DCE5E29@noak.vxo.telub.se> References: <2DC7EB42@noak.vxo.telub.se> <2q8m1v$rdh@ulowell.uml.edu> Subject: Re: Again: HyTime links in practice [Lloyd Rutledge] | The architectural form "nHyTime" was specified in the HyTime Draft | International Standard and was replaced by "sHyTime" in the final | version of the standard. Thus, of course, you should use "sHyTime". I have used a copy of the International Standard (not the draft) that I received from the SGML SIG Hyper, through Steven Newcomb, TechnoTeacher. The reason for me to use this one was that at the time I started to use the standard, the Swedish Standard Commision had a time scale for the delivery for about a year and a half (I have no idea of why, it does not usually take that long). On my copy it says "First Edition 1992-08-00", which is a bit cryptic, but it is the true standard, or at least, so say SIG Hyper. And I've got "nHyTime", not "sHyTime" in my copy. Also, In the change notes for the Catalog of HyTime Architectural Forms and HyTime SGML Specification, Version 2.0, June 28, 1993, by Charles F. Goldfarb, it says: || 3. CORRECTIONS IN THIS VERSION || || Although International Standards are subjected to extensive review and || proofreading, early users frequently discover errors. These are || reported to the Project Editor for eventual consideration by the || responsible standards committee, after which "corrigenda" may be || published to correct the errors officially. || || Several errors and ambiguities have already been reported; the proposed || corrections and clarifications (hereinafter simply "corrections") for || some of them are reflected in this document. While it must be stressed || that the corrections are not official (and theoretically might never || be), they are of such a nature that they are highly likely to be || accepted. Corrections affecting only the text of the standard (that || is, not the formal SGML specifications included in parts 4 and 5 of || this document) are preceded with a currency sign ($). || || Corrections added in this version of this Catalog are: :: || 06.4.2 nHyTime: change to "sHyTime" ("suppress HyTime element"); || only attributes of nested sHyTime elements are processed. Therefore, I'm not sure you are right, and I can't see that it is obvious that I have to use "sHyTime". I have not yet found any clear reference to what I'm supposed to use, and that's why I'm interested to know how the only HyTime engine available handles this. Peter -- Peter Bergstr\öm internet: pebe@telub.se Telub Inforum AB phone: +46 - 470 - 425 97 351 80 V\ÄXJ\Ö fax: +46 - 470 - 397 96 SWEDEN Newsgroups: comp.text.sgml Date: 08 May 1994 14:36:00 UT From: Peter Bergstrom \ Message-ID: <2DCE5E30@noak.vxo.telub.se> Subject: JPEG and MPEG notations Can somebody direct me to the correct syntax for the formal public identifier for the JPEG standard and MPEG standard? I believe they are currently both ISO-DIS standards, I've got these numbers: ISO DIS 10918 (JPEG) ISO DIS 11172 (MPEG) From this I imagine they could be specfied something like... \ \ ... but are there any publicly released declaration of these somewhere, which differ from my guess above? Thanks! Peter -- Peter Bergstr\öm internet: pebe@telub.se Telub Inforum AB phone: +46 - 470 - 425 97 351 80 V\ÄXJ\Ö fax: +46 - 470 - 397 96 SWEDEN Newsgroups: comp.text.sgml Date: 08 May 1994 14:58:00 UT From: Peter Bergstrom \ Message-ID: <2DCE5E36@noak.vxo.telub.se> References: <2DC91527@noak.vxo.telub.se> <2qc1je$g4o@search01.news.aol.com> Subject: Re: HyTime prolog [Peter Bergstrom] 1. They can be placed as the first thing in the Declaration Subset of | the DOCTYPE statement (which I believe is actually part of the base | document type declaration): : | 2. They can be placed in the SGML Declarations File, after the SGML | Declaration, of which APPINFO is the last parameter. This is | however a "grey zone" and I have no idea if any parser/HyTime | engine will ever see the declarations here. : | 3. They can be the very first thing of the DTD, but it seems to me | that then the "base document type declaration" has already started | (which might be the case even for alternative 1 above). [Eliot Kimber] | Your answer 1 is the only right answer. The intent for putting the | declarations in the SGML document proper, and not in the public part of | the DTD or in the SGML declaration itself, is that each instance may | have different requirements, and specifically, because HyTime | validation is done on an instance basis. Thanks Eliot, I will use this until I'm more satisfied with another answer. I still believe that what goes into the document type declaration subset is part of the base document type declaration, that's at least what I can figure out from the handbook, page 303 and 403. What confuses me on page 303 is that "other prolog" is possible before the base document type declaration, but I can't figure out where to put it, if not immediately after the SGML declarations (i.e., at the end of the same file = my original alternative 2). We're designing a DTD that must not be changed very much by each single user. It is not a meta-DTD. Small changes are allowed, which in practice means that the user uses a subset of the DTD. The HyTime support however, is always required in the same way. It is very important that the HyTime links are used for references in the way specified by the DTD. Thus, all instances will have the same HyTime support declarations. Is it then an error to place them as the first thing in the public DTD (=alternative 3), which is in fact the base DTD? | I don't know if any current tools enforce this requirement. Does anybody else know? Peter -- Peter Bergstr\öm internet: pebe@telub.se Telub Inforum AB phone: +46 - 470 - 425 97 351 80 V\ÄXJ\Ö fax: +46 - 470 - 397 96 SWEDEN Newsgroups: comp.text.sgml Date: 08 May 1994 20:48:02 UT From: Eliot Kimber \ Message-ID: <2qjj62$85k@search01.news.aol.com> References: <2DC91527@noak.vxo.telub.se> <2qc1je$g4o@search01.news.aol.com> Subject: Re: HyTime prolog [Peter Bergstrom] | 1. They can be placed as the first thing in the Declaration Subset of | the DOCTYPE statement (which I believe is actually part of the | base| document type declaration): | | \ | \ | \ | \ | \ | ]> [Eliot Kimber] | Your answer 1 is the only right answer. The intent for putting the | declarations in the SGML document proper, and not in the public part of | the DTD or in the SGML declaration itself, is that each instance may | have different requirements, and specifically, because HyTime | validation is done on an instance basis. This is in fact wrong. I had misremembered the meaning of "prolog", confusing the proper meaning with "internal DTD subset". The prolog of an SGML document is that part between the SGML Declaration (if present) and the document instance set. The HyTime standard says the declarations must occur in the document prolog before the base document type declaration (the "base document type declaration" is the first one specified (SGML documents may have multiple DTDs when the CURRENT feature is support)). Thus, the HyTime declarations can only go *before* the DOCTYPE declaration, e.g.,: \ \ \ \ \ \ Note that an SGML declaration cannot contain processing instructions. I apologize for any confusion I may have caused. -- \

W. Eliot Kimber (kimber@passage.com) Systems Analyst and HyTime Consultant Passage Systems, Inc., 9971 Quail Blvd., Suite 903, Austin TX 78758 +1 512 339 1400 465 Fairchild Dr., Suite 201, Mountain View, CA 94043, +1 415 390 0911 \

Newsgroups: comp.text.sgml Date: 09 May 1994 02:21:02 UT From: Torben Noerup Nielsen \ Organization: University of Hawaii Message-ID: \ Keywords: Equations, AAP Subject: Documentation For the last several weeks, I've been hunting around for documentation on a couple of standards. But without luck. I'm looking for documentation for the AAP DTD. I have heard that this is now an ISO standard. Does anyone have an ISO reference number for it? I also need documentation on ISO standard SGML equations. Could someone give me a reference for a document? The last one is CALS tables. I'm using the DOCBOOK DTD for some work and it doesn't contain a whole lot of documentation for the table part. All I'm after is some simple documentation on CALS tables so I can get them displayed using EBT's DynaText. I'd appreciate any help anyone can give me with this. The documents do not have to be free; I'm willing to pay the usual fees for standards documents. I just need to know who to call. Thanks, Torben Newsgroups: comp.text.sgml Date: 09 May 1994 07:04:00 UT From: Peter Bergstrom \ Message-ID: <2DCE63F1@noak.vxo.telub.se> References: <2DC91527@noak.vxo.telub.se> <2qc1je$g4o@search01.news.aol.com> <2qjj62$85k@search01.news.aol.com> Subject: Re: HyTime prolog [Eliot Kimber] | Thus, the HyTime declarations can only go *before* the DOCTYPE | declaration, e.g.,: | | \ | \ | \ | \ | \ | \ | | Note that an SGML declaration cannot contain processing instructions. | | I apologize for any confusion I may have caused. Thanks, Eliot! I think my answer to your last post is now overruled, but as you can see there, you did not cause any big confusion, not for me anyway. So the correct answer is my second alternative, then. Does anybody have any clue to if the SGML Parsers and HyTime engines out there recognizes these processing instructions if I put them at the end of my SGML Declarations file? Newsgroups: comp.text.sgml Date: 09 May 1994 07:55:42 UT From: Pamela Gennusa \ Message-ID: <0097E29946C99A00.232002B4@dpsl.UUCP> Subject: LAST CALL for SGML Europe '94 Year-in-Review News Requests Dear Users of SGML - A few weeks ago, we ran this "Call for Bits" and we have received a good response, much better than in past years. However, the majority of the responses have been from vendors rather than users. Now, that's OK. We were glad to receive every "BIT", but we know that vendors make a living in SGML because Users are USING it. So, let us know what your up to! C A L L F O R B I T S _________________________ By now a five-year tradition, the SGML Year in Review continues to open both the North American and European SGML series of conferences. As always, it attempts to represent exciting new projects and initiatives as well as provide updates on on-going work of interest to the conference attendees and to the readers of this newsgroup and the half-dozen newsletters that now reprint the presentation. The SGML Europe Conference is fast approaching, and with it, the Year in Review Call for Contributions. THIS IS YOUR INVITATION to sing your own praises, toot your own horn, give the world a paragraph or two (maximum!) to describe: 1 Relevant Standards Activity 2 Governmental (or EC/CE or other international) Initiatives 3 Major Industry Initiatives 4 New (since May '93) and Forthcoming Publications 5 Major or Especially Interesting SGML Implementations 6 Product Announcements (since May '93) 7 Any SGML Activity that fits in the "Miscellaneous" Category Of particular interest are updates on any activities reported in previous Year in Review collections. EVEN IF YOU ARE PLANNING TO SPEAK ABOUT THE RELEVANT ITEM AT SGML EUROPE, PLEASE ALSO SEND A FEW SENTENCES FOR THE Year in Review. ***NEW THIS SPRING*** For the first time ever, in keeping with the "news broadcast" style of presentation, the Year in Review will be accompanied by slides. You are invited to send relevant images for your item to Pam Gennusa's address below. The graphics should be in colour if possible and may depict individuals, legible computer screen images, or whatever else you think is appropriate. They may be funny. Please don't send colour slides; prints only. Vendors of SGML software are also invited to send announcements of software that has been released since last May. Please contact Pam or Yuri for more details. (We don't want to spoil the surprise for the rest of you.) As always, thanks in advance. Please ensure we have the relevant bits by May 12, 1994. Email submissions to: dpsl!plg@visionware.co.uk AND yuri@sq.com Mail graphics to: Pam Gennusa, DPSL, Database Publishing Systems Ltd 608 Delta Business Park Great Western Way Swindon, Wilts SN1 1LZ United Kingdom Ph: +44 793 512 515 Fax: +44 793 512 516 You may send news items without images. The images are a bonus. --------------------------------------------------------------------- Pamela Gennusa Yuri Rubinsky Chair, SGML Europe Chair, SGML '94 Director, President, SoftQuad Inc. Database Publishing Systems Ltd ### Newsgroups: comp.text.sgml Date: 09 May 1994 15:15:57 UT From: R Alexander Milowski \ Organization: University of Minnesota Message-ID: \ References: \ Subject: Re: Q: Is SGML what I'm looking for? [Frank Teusink] | I am looking for a way to transform a relational database of `famous | persons' (result of a hobby of me and a few others) to something that | is more flexible. [ stuff deleted ] | Via the HTML language, I arrived at SGML, which might be a solution to | my problem. Ideally, I'd like to transform the database to something | like | | \ | \

\Henry IV\ (Bolingbroke), \king of England\\

| | Son of \...\ and \...\. Born on | \...\ at ... | | blabla ... blabla ... | \ | | Is this possible within the framework of SGML? | Are there `SGML-databases' that can be used for something like this? | Where are they? Well, there are a *few* SGML databases but they are quite expensive. I might be worth your while, if you don't have the monies, to use a flat-file/index scenario. I have built SGML-systems in the past where indexes were used to speed up the access to SGML documents. Then, the documents where pulled into a application. In fact, this application was similar in the fact that it did not *need* SGML but SGML made it easier to handle and much easier to print. I would think that if you don't have a large number of people in your database you could use a flat-file/index structure. One other thing to consider is a translation from some RDB into SGML first. Then use the SGML for publishing purposes. | As an aside, I would be grateful for references to a good book on SGML. I swear by Goldfarb's SGML Handbook for programmer/analyst types. For authors, I couldn't make a recommendation since I'm not an author. ;) -- R. Alexander Milowski SGML Researcher milor001@maroon.tc.umn.edu Newsgroups: comp.text.sgml Date: 09 May 1994 15:24:00 UT From: Peter Bergstrom \ Message-ID: <2DCED95C@noak.vxo.telub.se> Subject: Data attribute troubles Question regarding data attributes. My DTD looks as follows: \ \ \ My instance looks like this: \ ]> \ Here comes the text of my document. \&MyRaster; \ This gives me an error in Omnimark, since I have already used the notation in an entity declaration, without the parser knowing it has got an attribute. I have both OMITTAG and SHORTTAG specified as YES in the SGML Declarations. It doesn't help specifying the attribute in the entity declaration. Neither does it help if I omit calling the entity in the text, the error occurs before that. Now, if I do the following it works: My DTD looks as follows: \ \ My instance looks like this: \ \ ]> \ Here comes the text of my document. \ With the ATTLIST spec coming before the entity declaration, it's perfectly all right. I can change the resolution etc. BUT, I don't want the ATTLIST being part of the document type declaration subset in this way. The attribute will always be the same for all instances using the DTD, and I can't see any reason for having to specify the attlist first, except that it doesn't work otherwise. Can I do this in another way? Am I really forced to put all data attributes in the declaration subset, like this? I have not found any real clues to this from the SGML Handbook. Thankful for any help. Peter -- Peter Bergstr\öm internet: pebe@telub.se Telub Inforum AB phone: +46 - 470 - 425 97 351 80 V\ÄXJ\Ö fax: +46 - 470 - 397 96 SWEDEN Newsgroups: comp.text.sgml Date: 09 May 1994 16:04:48 UT From: Ken Holman \ Message-ID: <9405092307.AA1636@notes.notes> References: <2DCE5E30@noak.vxo.telub.se> Subject: Re: JPEG and MPEG notations [Peter Bergstrom] | are there any publicly released declaration of these somewhere, which | differ from my guess above? Martin Bryan and I began an ad-hoc project within ISO/IEC JTC1/SC18/WG8 about three weeks ago at the latest international meeting in Arlington, Virginia. The draft document N1716 dated April 21, 1994 leads with the paragraph: "The following notation names and public identifiers are designed to make it easier for users to identify data encoded using ISO and other commonly used data interchange formats in document sets that reference data through ISO 9070 public text object identifiers." We are in the process of collecting suggestions for formal identifiers for both ISO and proprietary notations. The document is not _nearly_ complete, nor polished, as we have only been able to put in a couple hours of work on it during the week of meetings. Some of the identifiers were found to be incorrect when we returned to our resources at our own offices (resources that were not available at the meeting). However, we _did_ get started and produced enough of a document to which we could obtain comments. We were planning a more formal posting to this group to solicit input..... but thank you Peter for getting the ball rolling sooner. | From this I imagine they could be specfied something like... | | \ | | \ We were thinking of a more algorithmic approach when dealing with standards documents so that new standards could be addressed in such a way that people could create new identifiers without waiting for the notation to be included in a revision of this document. Our _draft_ idea for JPEG as published in N1716 will be changed to the following based on information received since the meeting to: \ Note the use of the title of the standard document as well as the complete standard identification (with a "-X" part number if applicable). Regarding MPEG, there are many parts to the standard that specify different components (audio, video, up to 1.5 mbit/s, etc), so a number of notations have been proposed. We are looking for input from SC29 regarding the correct titles of documents that are associated with specific data types. Whether this document becomes a TR (Technical Report) or not remains to be seen. We have solicited input from member bodies in the resolutions of the meeting document N1672: Recommendation 14. Collection of Notation Declarations: WG8 draws the attention of experts on data encoding to the information being collected in N1716, "Notation Declarations for data encoding standards". Our initial collection was gleaned from numerous publicly available and privately obtained DTDs. Needless to say the variety of declarations for the same formats underscored the usefulness for such a report. Please remember that we are also collecting notations for proprietary data types, eg: \ \ \ (note that these suggestions have not yet been commented on by their respective designers/owners) Any contributions offered will be greatly appreciated. We will compile a new list from contributions received. Contributions are sent to both Martin and myself when mailed to: notation@msl.isis.org Thanks in advance for all your input..................... Ken -- Martin Bryan The SGML Centre Phone/Fax: +44 452 714029 29 Oldbury Orchard Churchdown E-mail: mtbryan@cheltenham-he.ac.uk Glos GL3 2PU G. Ken Holman | Phone: +1 613 727 5696 x317 Vice President, R\&D | Fax: +1 613 727 9491 Microstar Software Ltd. | US: 1 800 267 9975 34 Colonnade Road North | BBS: +1 613 727 5272 Nepean Ontario CANADA K2E-7J6 | E-mail: gkholman@msl.isis.org Newsgroups: comp.text.sgml Date: 09 May 1994 18:11:55 UT From: Len Bullard \ References: \ \ Subject: Re: Q: Is SGML what I'm looking for? I concur with Milowski on reference to Goldfarb's SGML Handbook. For programmers, it is a good book. For author's it is too much generally. Practical SGML by Eric Van Herwijnen (spelled from memory) is the best book I've used. I also recommend Martin Bryan's SGML An Author's Guide as a companion. SoftQuad used to give out a nice brochure-like guide that explained the basics very well. I am an author, so that is from Mr Ed's own automated lips. Newsgroups: comp.text.sgml Date: 10 May 1994 15:00:17 UT From: Sam Wilmott \ Organization: Exoterica Corporation Message-ID: <1994May10.150017.16895@xgml.com> References: <2DCED95C@noak.vxo.telub.se> Subject: Re: Data attribute troubles [Peter Bergstrom] | My DTD looks as follows: | | \ | \ | \ | | My instance looks like this: | | \ | ]> | \ | Here comes the text of my document. | \&MyRaster; | \ | | This gives me an error in Omnimark, ... A paragraph of Clause 10.5 of ISO 8879, and the following note state: An attempt to redefine an entity is ignored, but is not an error. NOTE: This requirement allows an entity declaration in a document type declaration subset to take priority over a later declaration of the same entity in a public document type definition. The second text paragraph of Clause 11.1 of ISO 8879, and the following note state: The external identifier points to an entity that is referenced at the end of the declaration subset and is considered to be part (or all) of it. The effective document type definition is the combination of the declarations entered in the subset and the external ones. NOTE: A parameter entity declaration in the subset will have priority over another declaration for that entity in the external entity, as the external entity is parsed later. The first text paragraph of Clause 11.4 of ISO 8879 states: The notation name cannot be specified on another notation declaration in the same document type definition. The first text paragraph of Clause 11.4.1.2 of ISO 8879 (added by Amendment 1) states: The validity of a data attribute specification is determined by the attribute definition list associated with the data content notation. The attribute definition list must have been declared prior to this declaration. As a consequence of 11.1, the declaration(s) in the explicit DOCTYPE are processed first, and those in "mydoc1.dtd" second. As a consequence of 11.4.1.2, the ATTLIST for the attributes of the notation used by the entity "MyRaster" must occur prior to the ENTITY declaration. Which means that the ATTLIST cannot occur in "mydoc1.dtd" and the ENTITY declaration in the DOCTYPE part. Duplicating the NOTATION declaration in both parts does not help because of 11.4. The motivation for the order in which the DOCTYPE part and "mydoc1.dtd" are processed is provided by 10.5 and the note in 11.1: because the DOCTYPE part is parsed first, and all but the first ENTITY declaration for an entity are ignored, ENTITY declarations in the DOCTYPE part can be used to "parameterize" "mydoc1.dtd", and "mydoc1.dtd" can contain "default" entity declarations that are used if (prior) ones don't appear in the DOCTYPE part. However, this can only be done with entities. In particular, by 11.4, a second NOTATION declaration is in error. The solution to the difficulty described in Peter's message is to declare "mydoc1.dtd" as an entity in the DOCTYPE part and reference it prior to the ENTITY declaration, rather than referencing the entity from the DOCTYPE header, as follows: \ %mydoc1; \ ]> This forces "mydoc1.dtd" to be processed first and gets the NOTATION and ATTLIST declarations prior to the ENTITY declarations, where you want them to be. If you are having trouble with Omnimark, either as a customer with a support contract or as a potential customer evaluating our sampler, don't hesitate to contact our support department by phone (+1 613 722 1700), fax (+1 613 722 5706) or email (support@exoterica.com). Regards. -- Sam Wilmott, Exoterica Corp. sw@exoterica.com voice (613) 722-1700 fax (613) 722-5706 Product information: info@exoterica.com Newsgroups: comp.text.sgml Date: 10 May 1994 15:16:32 UT From: Len Bullard \ Message-ID: <9405101516.AA33052@source.asset.com> References: \ Subject: Re: How can I get information about SMSL? Send email to bmarkey@world.std.com or Permanent Wave Productions 18 Harris Avenue Shrewsbury, Mass 01545. At last accounting, Bryan Markey is the editor and can tell you about status. Newsgroups: comp.text.sgml Date: 10 May 1994 16:36:06 UT From: Robbin Wong \ Organization: Dep. Computer Sci. & Information Eng., Chiao Tung Univ., Taiwan, R.O.C Message-ID: \ Subject: How can I get information about SMSL? How can I get information about SMSL? The Standard Multimedia Scripting Language (SMSL) is a joint project between SC18/WG8, the developers of SGML and HyTime, and SC29/WG12, the Multimedia /Hypermedia Expert Group (MHEG) which started in 1992. If you know something about the Standard, please reply to me. Reply will be appreciate and thank you for your reading. Robbin Wong in Chinese Taiwan. Newsgroups: comp.text.sgml Date: 10 May 1994 18:14:38 UT From: Kerry Burton \ Message-ID: \ References: <2DCED95C@noak.vxo.telub.se> <1994May10.150017.16895@xgml.com> Subject: Re: Data attribute troubles [Peter Bergstrom] | My DTD looks as follows: | | \ | \ | \ | | My instance looks like this: | | \ | ]> | \ | Here comes the text of my document. | \&MyRaster; | \ | | This gives me an error in Omnimark... [Sam Wilmott] | The solution to the difficulty described in Peter's message is to | declare "mydoc1.dtd" as an entity in the DOCTYPE part and reference it | prior to the ENTITY declaration, rather than referencing the entity | from the DOCTYPE header, as follows: | | \ | %mydoc1; | \ | ]> | | This forces "mydoc1.dtd" to be processed first and gets the NOTATION | and ATTLIST declarations prior to the ENTITY declarations, where you | want them to be. Omnimark is definitely handling things according to the standard. I ran across the same question a year or two ago, and in my opinion this problem indicates a "hole" in the standard. Sam's solution works fine for Peter's particular situation, but what will Peter do when he wants to create "local" entity declarations to override those in his base DTD? Does he need to do something like the following? \ \ \ %mydoc1; \ ]> While certainly legal, this seems to go against the "flavor" suggested by the use of external declaration subsets. It's not Sam's or Omnimark's fault; there's no way around it! For some SGML editors which use precompiled DTDs, this does not work well at all. Sometimes the DTD to be precompiled is prohibited from having a DOCTYPE declaration. This allows the DTD to be referenced as an external declaration subset in a generated DOCTYPE declaration when the document is exported as SGML. It is my guess that this problem in the standard is the reason that some of these products have chosen to ignore data attributes altogether. The addition of data attributes in Amendment 1 fit in with the original standard ALMOST seamlessly, but this is one problem which either slipped by the reviewers or was deemed unresolvable and left that way. Are there any standard committee members out there who could fill us in on the reason for the incongruity, or what plans (if any) exist to resolve it? -- Kerry Burton Phone: (801) 222-7432 WordPerfect Corporation FAX: (801) 222-7477 MS E122 Email: kerry@wordperfect.com 1555 N. Technology Way "I speak for myself, not necessarily Orem, UT 84057 for WordPerfect (or Novell)" Newsgroups: comp.text.sgml Date: 10 May 1994 20:10:53 UT From: Robin Cover \ Organization: UT Arlington Message-ID: \ Subject: New Book on HyTime (DeRose & Durand) The new book of HyTime by Steve DeRose and David Durand is introduced in a file I found on Kluwer's public information disk. It contains a detailed Table of Contents listing, Foreword to the book (by Andy van Dam), ordering information, etc. Prepublication discount is offered through June 30th. I'll include below the TOC below, but the full file is available via anonymous FTP: ftp://ftp.std.com/ftp/Kluwer/books/hytime Robin Cover (hope Steve and David don't mind this publicity...) ------------------------------------------------------------------------ Making Hypermedia Work: A User's Guide to HyTime by Steve J. DeRose and David G. Durand CONTENTS Contents i Foreword xi Conventions xv Acknowledgments xvii PART I 1 Introduction 3 1.1 Purpose and scope of this book 6 1.2 What is HyTime for? 8 1.3 What can HyTime do? 9 1.4 How to use this book 10 2 Hypermedia Concepts 13 2.1 Introduction 13 2.2 Print examples of hyperdocuments 14 2.3 Documents and links 21 2.4 Open systems 27 3 Overview of SGML 35 3.1 What is SGML, really? 35 3.2 Parts of an SGML document 39 3.3 SGML basics 40 3.4 A sample SGML document 51 3.5 Optional SGML features 58 3.6 Formal public identifiers 61 4 HyTime Quick Start 65 4.1 Clink, the basic point-to-point link 65 4.2 Linking to destinations without IDs 67 4.3 HyTime locators 68 4.4 Graphic locations and geometry 70 4.5 Linking by searching 71 4.6 More complex link structures 72 4.7 HyTime and SGML element types 74 5 The Structure of HyTime 77 5.1 Basic HyTime concepts 77 5.2 Architectural forms 79 5.3 SGML issues in HyTime 90 5.4 HyTime's hyperdocument model 92 5.5 How a HyTime system fits together 94 5.6 HyTime declarations 97 5.7 The HyTime modules 99 PART II 6 Basic Measurement 103 6.1 Introduction to measurement 103 6.2 Geometric concepts 104 6.3 Architectural forms 105 7 Hyperlinks 109 7.1 Introduction 109 7.2 Basic Concepts 109 7.3 Architectural forms 114 7.4 Hyperlinks and webs 118 8 Locating Data Objects 121 8.1 Introduction 121 8.2 Naming, counting and querying 121 8.3 Named location addresses 123 8.4 Coordinate location addresses 131 8.5 Semantic locations 145 8.6 Creating robust pointers 151 8.7 Aggregates and spans 153 9 The HyTime Query Language 159 9.1 Purpose and basic capabilities 159 9.2 An example 160 9.3 HyQ syntax 162 9.4 The HyQ functions 163 9.5 HyQ and other query languages 186 10 Modifying a DTD for HyTime 189 10.1 Introduction 189 10.2 HTML and the World Wide Web 190 10.3 The DTD conversion 195 10.4 Evaluating a HyTime product 202 10.5 Summary 203 PART III 11 Lexical Types and Properties 207 11.1 Introduction 207 11.2 Introduction to lexical types 208 11.3 HyTime lextype type constraints 210 11.4 Defining lexical models 210 11.5 HyLex 213 11.6 Accessing public lexical types 220 11.7 HyTime properties 222 11.8 Property definition 227 11.9 Inherent property definitions 230 12 Extending HyTime 237 12.1 Introduction 237 12.2 HyQ extensions 238 12.3 Lexical types 240 12.4 Properties 241 12.5 Specific elements for media links 241 12.6 A protocol for hot spots on graphs 246 12.7 TEI 248 13 Advanced Measurement and Scheduling 253 13.1 Introduction 253 13.2 Units of measurement 254 13.3 Axes and finite coordinate spaces 257 13.4 Events and extents 260 13.5 Representing tables 267 13.6 Marker functions 271 13.7 Other features 273 14 The HyTime Base Module 275 14.1 Introduction 275 14.2 Entity trees 276 14.3 Base module option summaries 276 14.4 The all-id attribute form 280 14.5 The all-lex attribute form 284 14.6 The lexmodel architectural form 285 14.7 The lexord architectural form 285 14.8 The all-ref attribute form 287 14.9 The all-act 292 14.10 The any-dcn 293 14.11 The all-qual 294 APPENDIXES Appendix A HyTime Meta-DTD 295 A.1 Base module (clause 6) 296 A.2 Measurement (clause 7) 301 A.3 Location address (clause 8) 305 A.4 Hyperlinks (clause 9) 309 A.5 Scheduling (clause 10) 310 A.6 Rendition (clause 11) A.7 Useful types and notations 317 Measurement (annex A.5) 318 Appendix B Graphics Notations 321 Appendix C HyTime Reference Material 327 Bibliography 333 Glossary 341 Index 359 FOREWORD HyTime is the first official standard for describing the structure of time-based hypermedia documents. Its marketplace acceptance is yet to come. This book, presenting the first in-depth guide to the HyTime specification, both describes its key features and provides guidelines on how it is used. It has been written by two leading experts in the field who have had significant impact on its development and are experts in SGML, the prior standard that HyTime extends. [... continues] Newsgroups: comp.infosystems.www,comp.text.sgml,comp.lang.perl Date: 10 May 1994 23:58:43 UT From: Earl Hood \ Organization: Engineering, Convex Computer Corporation, Richardson, Tx USA Message-ID: <2qp73j$mbi@lovecraft.convex.com> Keywords: SGML, HTML, Web, Perl Summary: Hypertext navigation of SGML DTDs Subject: SGML DTD navigation with the Web (This message is being sent out to what I believe are interested parties.) Ever wondering if you can put a paragraph tag in a heading? What elements can contain a footnote? Can I put a graphic inside of a table? If you've asked yourself questions like above, then I have the program for you: dtd2html. Dtd2html is a Perl program that generates an HTML document (composed of several files) to allow hypertext navigation through an SGML DTD. Now one can easily find out what elements can contain what. Dtd2html determines what are the top-most elements in a DTD, and then allows a user to navigate down (or up) the DTD element structure via hyperlinks. Dtd2html also generates an HTML page containing a list of all elements so you can quickly jump to an element page you want. Dtd2html can also generate a page so one can see the overall element hierarchy defined by the DTD. Here's an example of what an element page looks like from the HTML DTD (This was generated thru Mosaic's save file option using formatted text; therefore, all the hyperlinks do not show up): --------------------------------------------------------------------------- HEAD **** The HEAD element contains all information about the document in general. It does not contain any text which is part of the document: this is in the BODY. Content ======= o base o isindex o link o nextid o title Tag Minimization Open Tag: REQUIRED Close Tag: REQUIRED See ATTRIBUTES See CONTENT DECLERATION Parent Elements =============== o html Goto top element list. Goto all element list. Goto tree. Goto HTML DTD main page. HTML DTD --------------------------------------------------------------------------- The element page contains hyperlinks to other element pages; a page describing the attributes of the element; a page giving the actual SGML syntax that defines the element; and links to the top element list, all element list, tree, and home pages. I've currently used dtd2html on the following DTDs: DocBook, OpenBook, Rainbow, HTML, HTML+, OSF Book I've added in the ability for dtd2html to add in descriptions of the elements in the DTD. This way, one can use dtd2html as way of documenting a DTD. Is anyone interested in such a program? I've created a tar file containing the program and the DTDs I've converted. The tar file also contains the following: dtdtree dtdtree outputs the content hierarchy tree (in ASCII) of SGML elements defined in a DTD. dtd.pl A Perl library that parses an SGML document type defintion (DTD) and builds up data structures containing the structural content of the DTD. Various convienance routines are defined to access the data structures. dtd2html and dtdtree utilize dtd.pl to do their work. If there is interest, I would like to see if anyone/site is willing to archive the stuff for distribution. Currently, our company is still working on geting a public Web server running, so it may be awhile before I can put it there. Following is the text file describing dtd2html: dtd2html ********* dtd2html is a Perl program that generates an HTML document (composed of several files) to allow hypertext navigation through an SGML DTD. Description ============ dtd2html generates various HTML files for hypertext navigation of a SGML DTD. The files generated are as follows: DTD-HOME.html This file is the home page of the HTML document. This file contains to the basic links to start navigating through the DTD. The name of this file can be changed with the -homename option. TOP-ELEM.html This file lists the top-most elements of the DTD, and contains the links to element pages describing each top-most element. The name of this file can be changed with the -topname option. ALL-ELEM.html This file contains a list of all elements defined in the DTD. This page allows quick access to any individual element description page. The name of this file can be changed with the -allname option. DTD-TREE.html (Optional) This file contains the content heierachy tree(s) of the top-most element(s) in the DTD. This file is only generated if the -tree option is specified during program invocation. The name of this file can be changed with the -treename option. element.html For each element defined in the DTD, an element description file is generated with a filename of the element name suffixed by ".html". element.attr.html For each element defined in the DTD, a file is generated describing the attributes defined for the element. element.cont.html For each element defined in the DTD, a file is generated listing the content model decleration of the element as declared in the DTD. Once all the files are generated, one needs only to create a link in the Web server being used to the DTD-HOME.html file. More information on the content of each file is in the HTML File Descriptions section. Usage ====== dtd2html is invoked from a Unix shell, with the following syntax: % dtd2html [options] filename filename is the SGML DTD to be parsed for generating the HTML files. The following is the list of options available: -allname filename Set the filename for file listing all elements in the DTD to filename. The default name is "ALL-ELEM.html". -descfile filename Use filename as the source for element descriptions in the DTD. If this argument is not specified, no description file is used. See Element Description File for more information. -dtdname string Set the name of the DTD to string. If not specified, dtd2html determines the name of the DTD by its filename with the extension stripped off. If reading from standard input, then this argument should be specified. Otherwise, "Unknown" is used. The string " DTD" will be appended to the name of the DTD. -elemlist Generate a blank element description file (see Element Description File). No HTML files are generated and all other options are ignored when this option is specified. -help Print out a terse description of all options available. No HTML files are generated and all other options are ignored when this option is specified. -homename filename Set the filename for the HTML home page for the DTD to filename. The default name is "DTD-HOME.html". -level # Set the prune level of the content hierachy tree to #. This option is only valid if -tree is specified. -mapfile filename Set the entity to map file to filename. The default is "map.txt". See DTDread_mapfile of package dtd for more information. The entity map file is only necessary if element/attribute markup are contained in multiply files, and the files are referenced by external parameter entities. dtd2html ignores general entities since they do not define document structure. -outdir path Set destination of generated HTML files to path. Defaults to the current working directory. -topname filename Set the filename for file listing the top-most elements in the DTD to filename. The default name is "TOP-ELEM.html". -tree Generate the content hierarchy of the top-most elements defined in the DTD. -treename filename Set the filename for file containing the content hierarchy tree(s) of the DTD to filename. The default name is "DTD-TREE.html". This option is only valid if -tree is specified. -treetop string Set the top-most elements to string. String is a comma separated list of elements that dtd2html should treat as the top-most elements when printing the content hierarchy tree(s). This option is only valid if -tree is specified. Normally, dtd2html will compute what are the top-most elements of the DTD. This option overrides that computation for the sake of of printing the content hierarchy trees. This option has no affect on the top-most element list presented "TOP-ELEM.html" (or the file specified by -topname). HTML File Descriptions ======================= All HTML files/pages generated contain hypertext links at the end of the page to the DTD-HOME, TOP-ELEM, ALL-ELEM, and DTD-TREE (if specified) files, unless stated otherwise. DTD-HOME +++++++++ This page is the root of the HTML document. It contains the links to the other main pages as described above. It also contains a brief paragraph mentioning that this document was created by dtd2html. This file may (should) be editted to add further description of the DTD, and/or contain links to other documents about the DTD. TOP-ELEM +++++++++ This page contains the list of all top-most elements defined in the DTD. A top-most element is defined as: An element which cannot be contained by another element or can be only contained by itself. ALL-ELEM +++++++++ This page contains a list of all elements defined in the DTD. DTD-TREE +++++++++ This page contains the content hierarchy tree(s) of the top-most elements of the DTD. See DTDprint_tree of package dtd for a detailed description about the content hierarchy tree(s). element ++++++++ element.html is the file describing the content of element. The element page is divided into the following sections: o element description (optional, see Element Description File). o Links to subelements broken into three parts: o Base content o Inclusions (if defined) o Exclusions (if defined) o Tag minimization (if defined). o Links to element's attribute (element.attr) and SGML decleration (element.cont) pages. o Links to all possible parent elements. o Links to main pages as described above. element.attr +++++++++++++ element.attr.html is the file describing the attributes of element. The element.attr page is divided into the following sections: o attribute(s) description (optional, see Element Description File). o List of attributes with possible values and default value. o Link back to element page. element.cont +++++++++++++ element.cont.html is the file giving the element content model decleration as it appears in the DTD. The element.cont page is divided into the following sections: o Base content decleration. o Inclusion content decleration (if defined). o Exclusion content decleration (if defined). o Link back to element page. Element Description File ========================= When dtd2html is invoked, a file containing descriptions of the various elements defined in the DTD may be specified (-descfile). These descriptions will appear in the element.html and element.attr.html files. The syntax of the description file is as follows: \ Description of element1 here. \

\ Description of element2 here. \

... The line \ signifies the beginning of the element description text. All text up to the next \ line or end-of-file is used as the element description. If element in the description file is suffixed by a "*", then the following description text will put in the element's attribute page. Otherwise, the description text appears in the element page. The description text can be any valid HTML markup. It is recommended to end each description with the \

tag so the description is separated from any other HTML markup generated by dtd2html. An \

(horizontal rule) tag is put after the description text by dtd2html. To get started with a description file for a DTD, one can use the -elemlist option to dtd2html to generate a file with all elements defined in the DTD with empty descriptions. See Also ========= dtdtree Generate content hierarchy trees of SGML elements. dtd.pl A Perl library to parse SGML DTDs. Earl Hood, ehood@convex.com -- Earl Hood | CONVEX Computer Corporation ehood@convex.com | 3000 Waterview Parkway Phone: (214) 497-4387 | P.O. Box 833851 FAX: (214) 497-4500 | Richardson, TX 75083-3851 Newsgroups: comp.text.sgml Date: 11 May 1994 09:58:40 UT From: Matt Timmermans \ Message-ID: <9405111702.AA0524@notes.notes> References: <2DCED95C@noak.vxo.telub.se> <1994May10.150017.16895@xgml.com> \ Subject: Re: Data attribute troubles [Kerry Burton] | \ | \ | \ | %mydoc1; | \ | ]> | | While certainly legal, this seems to go against the "flavor" suggested | by the use of external declaration subsets. It's not Sam's or | Omnimark's fault; there's no way around it! This gives me an opportunity to flog my favourite dead horse in a somewhat more eloquent way: The "flavor" suggested by the use of external declaration subsets is wrong. There is no difference between: \ and \ ... other declarations ... %thedtd; ]> The first example _seems_ to imply that the document somehow conforms to cow.dtd, but this is not the case. There is no way in SGML to declare conformance to a publicly known content model. This does not indicate a hole in the Standard. It indicates, rather, a false inferrence made by _everybody_ (myself included) who undertakes to learn SGML. \ -- Matt Timmermans | Phone: +1 613 727 5696 Microstar Software Ltd. | Fax: +1 613 727 9491 34 Colonnade Rd. North | BBS: +1 613 727 5272 Nepean Ontario CANADA K2E-7J6 | E-mail: mtimmerm@msl.isis.org Newsgroups: comp.infosystems.www,comp.text.sgml,comp.lang.perl Date: 11 May 1994 14:33:52 UT From: Joachim Schrod \ Organization: TH Darmstadt, FG Systemprogrammierung Message-ID: <2qqqcg$ptm@rs18.hrz.th-darmstadt.de> References: <2qp73j$mbi@lovecraft.convex.com> Keywords: SGML, HTML, Web, Perl Subject: Re: SGML DTD navigation with the Web [Earl Hood] | Ever wondering if you can put a paragraph tag in a heading? What | elements can contain a footnote? Can I put a graphic inside of a | table? | | If you've asked yourself questions like above, then I have the program | for you: dtd2html. And if you would tell us an ftp site, we would be most happy... :-) Cheers, Joachim -- Joachim Schrod Email: schrod@iti.informatik.th-darmstadt.de Computer Science Department Technical University of Darmstadt, Germany Newsgroups: comp.text.sgml Date: 11 May 1994 19:55:54 UT From: John Knauer \ Message-ID: <9405111956.AA18658@wpdis03.wpafb.af.mil> References: \ Keywords: Equations, AAP Subject: Re: Documentation [Torben Noerup Nielsen] | For the last several weeks, I've been hunting around for documentation | on a couple of standards. But without luck. | | I'm looking for documentation for the AAP DTD. I have heard that this | is now an ISO standard. Does anyone have an ISO reference number for | it? | | I also need documentation on ISO standard SGML equations. Could | someone give me a reference for a document? | | The last one is CALS tables. I'm using the DOCBOOK DTD for some work | and it doesn't contain a whole lot of documentation for the table part. | All I'm after is some simple documentation on CALS tables so I can get | them displayed using EBT's DynaText. | | I'd appreciate any help anyone can give me with this. The documents do | not have to be free; I'm willing to pay the usual fees for standards | documents. I just need to know who to call. In regards to your request for information regarding CALS tables, there is a file available via anonymous ftp which documents the table model in MIL-M-38784D. This specification is currently out for official coordination. Currently it is anticipated that this will be officially released around September. There were a number of improvements made in the table model in the current version of MIL-M-38784C AMENDMENT 3. The changes and the reasons for them are also contained in this file. This should help you understand the documentation for MIL-M-38784D as it applies to the DTD you are using. Whenever I see the phrase CALS tables I get a bit nervous because what is usually being referred to is the table model of Appendix A in MIL-M-28001B. This DTD has a number of problems. It was placed there only for illustration purposes and is duly noted as such, but still has been interpreted by many to be the CALS DTD. This interpretation is incorrect. This DTD is not being used by Air Force in any official capacity. To access this file login to WPAFTB1.WPAFB.AF.MIL as FTP and then type cd sgml. The file is table.doc. There are also a number of files relating to the Air Force CALS effort in this subdirectory. The file filelist.txt contains a listing of all the files in this subdirectory. -- John Knauer knauerj@wpdis03.wpafb.af.mil RJO Enterprises AFMC\\ENCS Newsgroups: comp.text.sgml Date: 11 May 1994 20:13:55 UT From: William L Trippe \ Organization: The World Public Access UNIX, Brookline, MA Message-ID: \ Keywords: programming Summary: Software programmers wanted in Boston Subject: Programming Positions in Boston WANTED: Senior Software Development Engineer WANTED: Software Development Engineer Boston, MA. Will Not Relocate. NO E-MAIL REPLIES, PLEASE!! InfoSoft International Inc. is a rapidly growing software developer that recently went public. (InfoSoft was formerly the Software Division of Houghton Mifflin Company.) Our present team of 100 professionals is predominantly software engineers and linguists. InfoSoft International has established a line of multilingual tools and reference works that are licensed to software developers and hardware manufacturers worldwide. Our platform-independent linguistic software products include: a spelling checker and corrector that works in 18 languages and dialects, a grammar correction system for American and British English, and a database retrieval system that can access electronic works such as dictionaries, thesauruses, and encyclopedias in many languages. We are also developing a new line of information management tools. All products are licensed to software OEMs who incorporate InfoSoft's software within their applications. Products are licensed as C source code and as such need to be portable to a wide range of computer platforms and operating systems including: MS-DOS, OS/2, MS Windows, Windows NT, Macintosh, VAX/VMS, and UNIX. The IntelliFinder is a research and development effort that is focused on information storage and retrieval technology. We work primarily with highly structured static reference works in many languages. =========================================================================== Senior Software Development Engineer: This position requires a forward thinking senior engineer with a focus on design and software architecture. The engineer must have experience with an entire product cycle, especially the design and specification stages. The engineer should have a demonstrated ability to document ideas and concepts. If possible we would like to see previous design documents. Previous experience with project leadership is desirable. The candidate should be strong in C, should have 7 or more years experience developing software, and should have recent experience developing for personal computers (PCs or Mac). Experience with cross platform development, data compression and multimedia are desirable. Knowledge of document mark-up and parsing (SGML) is also desirable. Knowledge of foreign languages is a plus. Initial tasks include specification of many of the new sub-systems. The engineer will help the project leader in reviewing all specification work done by other sources, and will be expected to review and contribute to all designs. The position will require code development and code review (for standards and performance). =========================================================================== Software Development Engineer: This is a junior coding and development position. The candidate should have strengths in coding new systems to specification. Also needed are skills in performance analysis, and integration of code. Experience with the C programming language in as many environments as possible is a plus. Experience with Windows NT, UNIX, or Macintosh or OS/2 is desired in addition to experience with the Windows development environment. Demonstrated coding ability and the ability to learn new concepts quickly is needed. Knowledge of foreign languages is a plus. =========================================================================== Summary: InfoSoft International Inc. offers an excellent opportunity for computer professionals to gain valuable experience in software development. We develop products for the latest computers, operating systems, and compilers, offering a unique opportunity to get hands-on experience with the newest development tools and systems. The pace of the group is quick and the work will be challenging. However, the opportunity for acquiring valuable experience in the software industry is tremendous and the work will be very rewarding. Please send or fax resume and salary requirement to: Pat Lewis InfoSoft International Inc. 222 Berkeley Street Boston, MA 02116 Fax #: 617-351-1115 NO E-MAIL REPLIES, PLEASE!! Newsgroups: comp.text.sgml Date: 11 May 1994 20:28:45 UT From: Henry Chung \ Organization: PANIX Public Access Internet and Unix, NYC Message-ID: <2qrf5t$5ig@panix.com> Keywords: SGML parser Subject: [Q] Whereabouts of a public domain parser? I am trying to locate the public domain parser. I had downloaded it once a while ago but have long since deleted it off my system. Recent archie searches have turned up nothing. Does anyone know where I might find it? Any help would be much appreciated. Thanks. -- Henry Chung (hchung@panix.com) Technical Director US Lynx Newsgroups: comp.infosystems.www,comp.text.sgml,comp.lang.perl Date: 11 May 1994 22:01:26 UT From: Earl Hood \ Organization: Engineering, Convex Computer Corporation, Richardson, Tx USA Message-ID: <2qrkjm$b0h@lovecraft.convex.com> Summary: ftp sites for dtd2html Subject: DTD2HTML NOW AVAILABLE (was Re: SGML DTD navigation ...) dtd2html is now available (or will be soon) at the following locations: ftp://ftp.law.indiana.edu/pub/www/ ftp://ftp.funet.fi/pub/networking/services/www/misc/ http://caesar.cs.uiowa.edu/~jollie/dtd2html/ ftp://ftp.ifi.uio.no//pub/SGML/DTD2HTML/ and at ftp.utdallas.edu once a Web page is written for it. More may get added over time. I'd like to setup a standard mechanism for providing later versions (if any) to published. There are 2 files in the distribution: dtd2html.tar Tar file containing the distribution of dtd2html and other related tools and files. dtd2html.examples.tar Tar file of DTDs processed with dtd2html, and sample output from dtdtree. Thanks to those willing to archive the files. All locations listed above are subject to change without notice. If there are any bug/comments/suggestions about dtd2html and related tools, please drop me a message. --ewh -- Earl Hood | CONVEX Computer Corporation ehood@convex.com | 3000 Waterview Parkway Phone: (214) 497-4387 | P.O. Box 833851 FAX: (214) 497-4500 | Richardson, TX 75083-3851 Newsgroups: comp.text.sgml Date: 12 May 1994 13:04:28 UT From: Len Bullard \ Message-ID: <9405121304.AA16979@source.asset.com> References: \ <9405111956.AA18658@wpdis03.wpafb.af.mil> Subject: Re: Documentation Isn't MIL-M-28001B the DTD that is being quoted for contracting purposes? If there is confusion, it is widespread and spreading faster. I've received several phone calls lately in which I'm asked about tools to support production using the "CALS DTD per MIL-M-28001". If this is not the DTD to which they are referring, a very strong message should come out of the CALS Policy Office, the ISG EPC or someone. Does the Air Force provide or recommend a FOSI for the 38784 DTD? Len Bullard Unisys Huntsville Newsgroups: comp.text.sgml Date: 12 May 1994 14:35:14 UT From: Debby Young \ Message-ID: <9405121435.AA02475@tecnet1.jcte.jcs.mil> References: \ <9405111956.AA18658@wpdis03.wpafb.af.mil> <9405121304.AA16979@source.asset.com> Subject: Re: Documentation [Len Bullard] | Isn't MIL-M-28001B the DTD that is being quoted for contracting purposes? I don't believe that the 38784 DTD that was attached to 28001B is the only DTD being quoted for contracting purposes. I know that the air side of the Navy uses another format, work package, for which a DTD and FOSI is being finalized. | If there is confusion, it is widespread and spreading faster. I've | received several phone calls lately in which I'm asked about tools to | support production using the "CALS DTD per MIL-M-28001". If this is | not the DTD to which they are referring, a very strong message should | come out of the CALS Policy Office, the ISG EPC or someone. From what I have gathered in my travels, the DTD attached 38784 was never functionally tested to work, and the JCALS folks proved that it would not support their requirements. There is going to be a revision coming to 28001; I hear that because of JCALS findings that the priority to get the standard updated has been elevated. | Does the Air Force provide or recommend a FOSI for the 38784 DTD? I don't know about the Air Force providing a FOSI (I haven't found one yet) but the Naval Air community will be finalizing the FOSI for 38784C for our use by the end of June. The Naval Air Electronic TM Team has also produced an IDEF model which defines the management process for DTDs and FOSIs, including configuration control, improving, and testing. When the FOSI is finalized, we will place it in the public domain. -- Debby Young Naval Air Warfare Center Weapons Divsion Point Mugu, CA young@tecnet1.jcte.jcs.mil (805) 484-6576 Newsgroups: comp.text.sgml Date: 12 May 1994 14:57:58 UT From: Len Bullard \ Message-ID: <9405121457.AA48131@source.asset.com> Subject: Treeloc in HyTime In the example given in ISO10744, the marklist is used to locate third element of the second element in the second level: \1 2 3\ where \ contains \ and \. \ is root (1) and \ is the second element of the second level (2). \ has three \